Epidemiologic studies of senile dementia are needed in order to determine if the disease occurs in specific patterns by age, sex, race or other demographic and social characteristics and if such patterns would provide etiologic clues. This project will characterize persons within a community based, non-institutionalized sample who appear to have senile dementia and/or depression or to be at high risk of having these disorders. This sample will be compared with an unaffected group. Both samples will be followed over a three year period to discover risk factors and etiologic events.